Re: [PATCH v2 2/2] s390/kexec: fix kmemleak

From: Baoquan He <bhe@redhat.com>
To: Heiko Carstens <hca@linux.ibm.com>
Cc: kernel test robot <lkp@intel.com>,
	linux-kernel@vger.kernel.org, kbuild-all@lists.01.org,
	linux-s390@vger.kernel.org, kexec@lists.infradead.org,
	prudo@redhat.com
Subject: Re: [PATCH v2 2/2] s390/kexec: fix kmemleak
Date: Fri, 19 Nov 2021 10:35:31 +0800	[thread overview]
Message-ID: <20211119023531.GB12774@MiWiFi-R3L-srv> (raw)
In-Reply-To: <YZYUi9XMyjtpGmAQ@osiris>

On 11/18/21 at 09:53am, Heiko Carstens wrote:
> On Thu, Nov 18, 2021 at 03:13:27PM +0800, Baoquan He wrote:
> > On 11/18/21 at 05:46am, kernel test robot wrote:
> > >    arch/s390/kernel/machine_kexec_file.c: In function 'arch_kimage_file_post_load_cleanup':
> > > >> arch/s390/kernel/machine_kexec_file.c:332:9: error: implicit declaration of function 'kvfree'; did you mean 'vfree'? [-Werror=implicit-function-declaration]
> > >      332 |         kvfree(image->arch.ipl_buf);
> > >          |         ^~~~~~
> > >          |         vfree
> > 
> > OK, kvfree is not wrong, seems vfree is more appropriate since it's
> > clear the ipl_buf is allocated with zvalloc() in ipl_report_finish().
> > 
> > Hi Heiko,
> > 
> > Could you help modify the code in your tree or append below patch to
> > mute the lkp complaint? Sorry for the inconvenience.
> ...
> >  arch/s390/kernel/machine_kexec_file.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/s390/kernel/machine_kexec_file.c b/arch/s390/kernel/machine_kexec_file.c
> > index 7f51837e9bc2..351a7ff69a43 100644
> > --- a/arch/s390/kernel/machine_kexec_file.c
> > +++ b/arch/s390/kernel/machine_kexec_file.c
> > @@ -329,7 +329,7 @@ int arch_kexec_apply_relocations_add(struct purgatory_info *pi,
> >  
> >  int arch_kimage_file_post_load_cleanup(struct kimage *image)
> >  {
> > -	kvfree(image->arch.ipl_buf);
> > +	vfree(image->arch.ipl_buf);
> 
> The problem reported above indicates that slab.h was not
> included. With your patch, while it fixes the problem for this
> particular configuration, this requires vmalloc.h to be included.

Indeed, thanks.

> 
> I'll merge your patch and add the missing include as well.
> 